ArtificialGuyBR

Home / Blog / Training LoRAs

ClayAnimation Redmond QWENIMAGE LoRA

LoRA for Qwen Image that generates clay animation-style images with soft, tactile 3D characters and warm, storybook lighting.

1 sources cited Training LoRAs

On this page

  1. What it does
  2. Base model and training
  3. Trigger words and prompts
  4. How to use
  5. Examples
  6. Common errors and fixes
  7. FAQ
  8. Sources

ClayAnimation Redmond QWENIMAGE LoRA

A LoRA for Qwen Image that applies a consistent clay animation style to generated images.

What it does

ClayAnimation Redmond QWENIMAGE adapts Qwen/Qwen-Image-2512 to produce images that look like high-quality clay animation. The visual signature is soft, rounded geometry; matte, tactile surfaces; and warm, diffused lighting. Characters read as physical figurines rather than glossy 3D renders.

The style works across subjects. Sample outputs show a family portrait on a couch, an underwater scene with a purple octopus and tropical fish, and a baker in a rustic bakery surrounded by bread. In each case, the clay texture and color palette remain consistent without requiring extensive prompt engineering.

Base model and training

This LoRA targets Qwen/Qwen-Image-2512. The model card states it was trained on clay animation-style images with an emphasis on detail and consistency. Weights are provided in Safetensors format. No exact dataset size or training duration is listed on the model card.

Trigger words and prompts

Use these trigger words to activate the style:

Pair them with subject descriptions as you normally would. The LoRA is intended to sit on top of standard Qwen Image prompts rather than replace them.

How to use

  1. Load Qwen/Qwen-Image-2512 as the base model.
  2. Apply the ClayAnimation Redmond QWENIMAGE LoRA adapter.
  3. Include Clay animation and Clay in the prompt.
  4. Generate in ComfyUI for the author-recommended workflow.

No specific weight or CFG values are documented on the model card. Start with typical LoRA strengths and adjust if the style overwhelms or under-performs.

Examples

Family scene in clay animation style

Underwater clay animation scene with an octopus and fish

Baker in a bakery rendered in clay animation style

Common errors and fixes

Error | Cause | Fix | Source

|---|---|---|---

Style does not appear | Trigger words missing or low LoRA weight | Add 'Clay animation' and 'Clay'; increase LoRA strength gradually | HF model card Inconsistent character design | Prompt too open-ended | Add character descriptors and scene constraints | Image inspection Model not loading | Wrong base model | Use Qwen/Qwen-Image-2512, not SD 1.5 or other checkpoints | HF model card

FAQ

Q: Can I use this with AUTOMATIC1111? A: The model card explicitly recommends ComfyUI. It may work elsewhere, but results are not guaranteed.

Q: Is this the same as the SD 1.5 ClayAnimationRedmond? A: No. This LoRA targets Qwen Image. Older ClayAnimationRedmond releases target Stable Diffusion 1.5.

Q: What license covers this model? A: Apache 2.0.

Sources